Fetching the paper…

Low-Precision Reinforcement Learning: Running Soft Actor-Critic in Half Precision · Around